You wouldn’t design a hospital yourself.
You would never knock through a wall without asking which ones hold the roof up. You do it to your business every quarter.
Nobody would let you design a hospital. Somebody would stop you long before the first wall went up, and rightly, because the stakes are too high for good intentions to be enough.
Nobody stops you designing your own business.
You know it needs goals, roles, a rhythm of meetings, the same way you know a house needs walls, a roof, and somewhere for the water to go. That is not architecture. That is a materials list. Most businesses are running on one, assembled by someone doing this for the first time, because nobody made them wait for someone who was not.
The reflex, and why it fails
Ask a leader what their business needs and you get the materials list back immediately. A strategy. Some goals. The right people in the right seats. All true, and none of it is the hard part.
The hard part is knowing which of those things is load-bearing.
A wall looks like a wall. Some are just standing there, dividing one room from another, harmless to remove. Others are holding the roof up, and you cannot tell which by looking. You have to know what it is actually connected to.
Most businesses have exactly this problem, and almost nobody goes looking for it, because an incentive looks even less structural than a wall does.
The wall that was never just a wall
Here is what that looks like in practice.
A sales team is missing its number, through no fault you could point a finger at, through a product that has fallen behind. Sales does the rational thing: discounts harder to close what it can, and starts quietly promising things the roadmap never actually committed to, because that is what closing the gap requires.
Product and engineering feel that promise land on their desk as a commitment they never agreed to. They did not cause the miss. They are catching the consequence of it anyway. So they disengage from the one thing they can still control, which is quality, on the entirely reasonable logic that caring about it clearly did not change anything last time.
Which makes the product worse. Which is what sales was trying to fix in the first place.
Nobody in that chain is behaving badly. Everyone is responding sensibly to what they can see from where they are standing. The wall that came down was invisible before anyone touched it: an incentive on one side of the business, quietly load-bearing for a team on the other side of it, with nobody having drawn the line between them because it was never on anyone’s plan.
There are no drawings
A building has drawings. That is exactly why an architect can walk in years later, look at a wall, and tell you whether it is holding the roof up. The knowledge outlived the people who built it.
Your business has no drawings.
Nobody wrote down why the approval threshold was set at that number. Or what the incentive was originally protecting. Or which team quietly depends on a process that looks decorative from the outside. The reasoning existed once, in somebody’s head, and that person has moved on.
So when you look at your own business and cannot tell which walls are structural, that is not a failure of intelligence. The drawings were never made. We have written about that elsewhere as organisational amnesia, and this is what it costs you in practice: not forgotten facts, but the inability to tell what your own business is standing on.
It is also why the fix is more concrete than it sounds. A metric model is a structural drawing. It is the thing that states what is connected to what, which is precisely the knowledge that separates a wall you can remove from a wall that is holding the roof up.
Somebody else’s collapse
Some of it is written in someone else’s mistake.
Building codes exist because a building fell down once, somewhere, and the industry decided nobody else should find that out the hard way. Most of what a serious operating system insists on, decision rights, span of control, a real cadence for reviewing risk, was learned exactly the same way, by another business finding out first. A founder building this for the first time has simply never had reason to know the lesson exists.
I have removed a load-bearing wall. More than once. The one I still think about was an approval step I took out because it was slowing everything down, and it turned out to be the only place two functions ever spoke to each other before a decision was made. Nothing failed for a month, which is the part worth telling.
The crack that shows up later
And some of it does not show up for years.
A foundation poured slightly wrong does not fail on day one. A design that was wrong at the start rarely announces itself at the three month mark. It shows up later, as a hairline crack, long after the cost of fixing it has multiplied past what pouring it correctly would have cost.
Why this is not quite architecture
Here is where the comparison actually breaks, and it breaks in the one place that matters most.
A wall does not care what the people living under it think of it. It holds the roof up whether they love the house or resent it, because it answers to physics, not opinion.
Your business does not answer to physics. It answers to belief.
An incentive, a process, a goal cascade, holds up for exactly as long as the people inside it keep believing it is real and keep acting as though it is. The day belief goes, the structure is still standing on the org chart and has already fallen down in practice, and nobody needed to touch a single wall for that to happen.
Which means a business can pass every inspection on the org chart and already be structurally unsound. The reporting line exists. The process is documented. And the people inside it stopped believing it was real two quarters ago, so it is holding nothing up. You will not see that in a diagram, and it will not announce itself until something leans on it.
You get to choose
Almost nobody designed their business on purpose. It accreted. A threshold set when cash was tight. A promotion nobody explained. A meeting format someone inherited and never questioned. None of it was chosen as a system, and all of it is teaching something, whether or not you meant it to.
You would not skip the architect for a building you actually cared about. Your business deserves the same question, and it is the harder one to ask, because nobody is going to stop you skipping it.
What in your business is actually holding the roof up, and does anyone still know why it is there?
This is the part of the work that never fits on a slide: finding what is actually load-bearing, and redesigning it properly before it cracks, rather than after.
In practice that means reading the artefacts, the goals, the incentives, the last ten promotions, the processes nobody can quite explain, alongside anonymous conversations with the people who work inside them. The artefacts tell you what was built. The conversations tell you what it is holding up.
Talk it through with Matt →Ten questions, three minutes, scored in your browser. It reads the system you have built, not the people in it.
Score today’s design →A UK-based strategy and OKR consultant and two-time SaaS founder with a venture-backed exit, Matt turns strategy into execution for teams scaling from tens to thousands. He co-founded ZOKRI in 2018, having previously co-founded Linkdex, a venture-backed enterprise SaaS platform he led to a trade sale. He writes the methodology behind these notes.